FR2535490A1 - COMPUTER SYSTEM FOR PROCESSING VOICE DATA - Google Patents

Abstract

THE INVENTION DESCRIBES A SYSTEM ALLOWING TO VIEW ON A SCREEN 31 THE STRUCTURE OF A VOICE MESSAGE, TO THE MEASURE OF ITS DICTATION, WITH THE POSSIBILITY OF INCORPORATING NOTES OR ANY MARKS BY A KEYPAD 16. THE VOICE MESSAGE FROM IT. FROM A MICROPHONE 50 TRANSIT BY AN ANALOGUE-DIGITAL CONVERTER 14 AND A SERIAL-PARALLEL CONVERTER 30 TO BE STORED IN A MEMORY 22 UNDER THE CONTROL OF A PROCESSOR 26. THE MESSAGE STORED IN MEMORY CAN BE MODIFIED AT WILL BY ADDING, DELETION OR REPLACEMENT OF SEGMENTS OF THIS MESSAGE, AND IT MAY BE REPRODUCED BY A LOUDSPEAKER 56.

Description

COMPUTER SYSTEM FOR

PROCESSING OF VOICE DATA

  The present invention relates to the editing of voice data. The subject of the present invention is an information processing system comprising means for acquiring a continuous signal intended for the acquisition of a continuously varying electrical signal corresponding to a voice message, means for digitizing intended for digitizing the signal. continuously variable electric to produce discrete voice data corresponding to audible quality

  voice message, data acquisition means available

  peaks intended for the acquisition of discrete data corresponding to alphanumeric characters, means for the acquisition of discrete signals intended for the acquisition of discrete signals comprising editing and command orders, a memory intended for recording data in discrete form, display means intended for

  produce a visualization, and a processor, all of these

  elements being functionally interconnected by conduits

  order tutors and data transfer channels,

  with a program for operating the processor

  stored in memory so that the processor controls the operation of the system to: record data

  discrete voice in memory, simultaneously with the acquisition

  voicemail, save character data

  res in memory simultaneously with character input,

  establish a specific sequence record in memory

  for a unified order of voice message and character data, view a sequence of symbol marks for voice data and character marks, each mark

25354 O

  symbol representing a predetermined increment of acquired voice message and each character mark corresponding to one of the characters entered, the displayed sequence corresponding to the sequence present in the recording, and revise the sequence recording, under the dependence

  of edit orders introduced, to take account of changes

  editing elements in the order of voice and character data

mothers.

  The invention may further involve a pro-

  operating gram such that the processor further controls the functioning of the system for: reacting to predetermined discrete signals acquired simultaneously with

  acquisition of the voice message, to indicate in the recording

  sequence recording the point at which each predetermined discrete signal was acquired; present on the display a distinctive indication of the moment at which each of these simultaneously acquired signals was acquired, in relation to other elements of the voice data; establish in memory a

  pointer defining a pointer position in the sequence

  this data, present in the display a visible mark corresponding to the pointer position; move, depending on acquired input signals, the position

  pointer defined in sequence and move this posi-

  correspondingly in the visualization; producing, under the dependence of acquired input signals, a continuously varying audio signal which corresponds to the discrete voice data recorded in memory, the production of this signal starting at a point in the sequence of voice data which corresponds to the position of defined pointer, as defined at the time, and following

  the order which is defined at the time considered in the record-

  sequence ment; and advance the pointer in the voice message data, in correspondence with the progress of the production of the audio signal The invention can also make

  intervene a circuit intended to detect the acquisition activity

25354 O

  audio information, and stop recording voice message data to memory if there is no activity. The invention provides an author with a visible graphic representation of the structure of the text he dictates, with

  indications that it may insert concerning paragraphs

  phes or other functional divisions It allows an author to edit with great flexibility the text he

  dictates, by displacement, deletion, insertion and reproduction

  tion, while the visual presentation helps him to follow the editing operations and to localize the point at which he must carry out editing revisions The invention also allows the author to introduce by a keyboard notes and intermediate instructions in its recording

dictated.

  The invention will be better understood on reading

  the following description of an embodiment given

  by way of nonlimiting example The rest of the description

  refers to the attached drawing which represents in the form of

  block diagram of a system according to the invention.

  A compliant voice data editing system 10

  me to the invention includes connections 12 intended for

  acquire and provide a continuously varying electrical signal

  naked corresponding to a voice message An acquired signal can

  be obtained from a microphone 50 or a telephone line

  phonic 52, via an interface circuit 54 as shown in the representation of the figure, or in other ways The signal supplied can be used to attack a

  speaker 56 as shown, or in other ways

  res Connections 12 are connected to a converter

  analog-to-digital 14, which performs bidi-

  The converter 14 is itself connected to a serial-parallel converter 30 operating in both directions. An audio sensor 28 is connected to the connections 12 and it emits a control signal which indicates whether there is a

253549 O

  activity on the voice message acquisition channel The system 10 also comprises a display unit 31 which may advantageously include a cathode screen, and a keyboard 16, which has a section 18 for entering alphanumeric characters and a section 20 for entering 'Entrance

  edit and control signals.

  The system 10 also includes a processor 26, which can be the model Z 80 manufactured by Zilog, and a memory 22 which is intended to record data in the form of bits and which has a section 24 in which

  an operating program is saved.

  elements of the system described above are interconnected by a data bus 58, an address bus 60 and control conductors 62, as shown in the figure All the elements of the

  system-10 described above are conventional items available

  commercially available, and how to interconnect them

  ter is well known to those skilled in the art.

  The operating program of the voice message editor which is saved in memory controls the operation of the system, in association with the processor.

  26, in the performance of all the functions of the edi-

  voicemail message When an author using the system speaks into a microphone, the voicemail acquired by the system

  in the form of an analog signal is digitized and intro-

  in memory in discrete form Simultaneously, a

  representation of the voice message using a series of symbols

  the vocals, each of them representing a second of message

  voice, is produced and displayed on the cathode screen.

  During voice pauses, data entry is

  removed to avoid wasted storage capacity

  re Simultaneously with the dictation, the author can introduce by the keyboard interrupt signals which produce

  memory pointers indicating when-introduce it-

  tion was made, in the recording of data and which cause the visualization of successive vocal symbols from 25354 t O

  of the next visualization line, which simulates a

  pure corresponding to a paragraph Simultaneously, the system produces a number carried in margin to allow to easily identify the cut The author can also interrupt the dictation with a signal entered on the keyboard and enter an alphanumeric text by the keyboard This text is entered

  in memory and is displayed on the display cathode screen

reading.

  The system operating under the control of the pro-

  gram keeps a record indicating a uniform sequence

  based on voice data, text data and cutoff indications The order of this sequence is initially

  the time order in which the system acquired the data.

  The system also produces a memory pointer which indicates

  as a pointer position in the data sequence A cursor mark is presented on the visualization, at a corresponding position The author can manipulate this pointer and this linked cursor mark to designate any particular point in the data sequence unified Using the cursor and keyboard edit signals, including "insert", "delete", "replace", "move" and "copy", the author can accomplish all of these

  editing functions, applying them without having to determine

  ner if the data is voice, text or

  brands The presentation made on the display device

  sation shows all editing changes as they are made. Using the cursor and keyboard input signals, the author can also play the voice message with any device.

connected audio.

  A more detailed description of the

  operation is made below, A voice message editor operating program is recorded in the memory 22 and, in association

  with processor 26, it controls the operation of the system

253549 O

  me in the performance of all the functions of the edi-

  voice message program The message editor program

  voice uses a program queue, and sub-

  programs called by the voice message editor are first introduced into the program queue, and they are executed thereafter when the processor reaches their level With such a queue, a program of

  interrupt management introduces a subpro-

  gram intended for the treatment of the interruption, then it reva-

  immediately interrupts and returns to the main program Subprograms are entered into the queue and are taken over by the processor when the latter is available to do so A program queue module contains subroutines intended for manipulate

  the voice message editor program queue.

  These are:

  INIT $ FILE $ PROG: Initializes the program queue

  my. INTR $ FILE $ PROG: Introduces a procedure address and an address parameter in the queue

of programs.

  EXEC $ FILE $ PROG: Checks if a procedure / parameter pair is in the queue If so, it calls the procedure by transmitting it

the unique address setting.

  The main program of the voice message editor is very simple, because of the program queue of

  the voice message editor The main editor program

  Voice message guardian performs two functions: 1) He calls

  an initialization program, INIT $ EDITOR $ VOCAL, for ini-

  tialize all data structures and hardware input / output devices used by the voice message editor; 2) It then loops indefinitely, calling

  EXEC $ FILE $ PROG to execute any other sub-

  programs in the program queue If the user indicates that he wishes to exit from the voice message editor, for example, the EXIT $ EDITOR procedure is 25354 to

  added to the program queue The process

  seur calls this program as soon as he can, which returns the voice message editor to the application that

called him.

  The previous description shows that a

  after the voice message editor has been called up and initiated

  read the variables and the material, it simply loops while waiting for something to appear in the queue

  waiting for programs We use inter-

  break to place something in this queue.

  Interrupt procedures are executed when a hardware interrupt occurs. When this happens, the processor invalidates the interrupts, it enters the

  stacks the current program counter and proceeds to a procedure

  re intended for the management of the interruption.

  The voice message editor works in the mode

  Interrupt 2 of the Z 80, and it receives pro

  from the following devices, given in order of priority

  rity of interruptions: 1 CTC O channel 2 CTC channel 1 3 CTC channel 2 4 CTC channel 3 Block count: this channel produces an interrupt when the audio material has just finished recording or

  reproduction of an information buffer

digitized audio.

  Telephone Ring: This channel produces an interruption each time the telephone rings. Press a key: the voice message editor programs this channel so as to produce an interruption each time a key press is received,

  Rythmeur: the voice message editor pro-

  gram this channel so that it produces a

  interruption every 10 ms (0.010 sec-

of).

  The addresses of interrupt management programs

  tion for the above devices are in a table

  interrupt vectors, in memory When any one

  that of the above devices produces an interrupt,

  the corresponding address in the interrupt vector table

tion is called.

  The interruption management programs of the edi-

  voice message tutors are found in two modules: the modu-

  the interrupts and the input / output management program module. The interrupt module consists only of a group of programs at assembly level, one for one

  each device producing interruptions All these pro-

  grams save the registers to the stack, what we

  calls a PLM procedure and then they restore the regis-

  very, validate the interruptions and return The programs-

  management messages are as follows: audio: CTC 0 channel management program, calls the

PLM INTERRUPT $ AUDIO procedure.

  ringing: CTC 1 channel management program, calls the

INTERRUPT $ RINGER procedure.

  GES Te TOUCHE: CTC 2 channel management program, performs an IN (00) operation to determine the key

  which has been pressed, save this information

  tion in a TOUCHEBRUTE variable, calls the

PLM procedure KEY $ OBTAINED.

  timer: CTC 3 channel management program, calls the

PLM RYTHMEUR $ TEN $ MS procedure.

  The input / output management program module contains PLM procedures which carry out most of the interrupt management I 1 also contains

  some other miscellaneous programs Interrupt programs

  tion are briefly described below:

  INTERRUPT $ RINGER: Entered in the production queue

  grams a procedure that will introduce the

  message: "Your phone rings,

go to TAB ".

  $ O BUTTENED KEY: Its characteristic function is simply to introduce the PROCESSING $ TOUCHIE procedure in the program queue. It is TREATMENT $ KEY which actually performs the management of

pressing a key.

  RHYTHMER $ TEN $ MS: Calls other PLM procedures that perform periodic checks

relating to certain conditions.

  Almost all the functions of the message editor

25354.0

  voice messages are triggered when the user presses a key The voice message editor uses a system based

  on a table to determine the procedure to be called in response

  when pressing a given key.

  The workstation keys are divided into 16 different categories A number from O to 15 is assigned to each category No key can appear in more than one category The category numbers and the keys

  each category is shown below.

Category number

9

Description

  category "record button" category "stop button" category "play / stop button" category "cursor" category "passage" category "number" category "text" back category "category category tag" category "spacing" brand "

"renumbered-

  "edition" category "execution" category "cancellation" category "help button" category "telephone button" category "invalid button" Keys

INSERTION

STOP Spacebar,

(INITIAL POSITION)

  North cursor, East cursor, South cursor, West cursor

GO TO THE NEXT PAGE-

TE 0 to 9 A Z, a z, comma,

dot l * ó * & () -

= + l l ;: '"/? z

Backspace Touch

re

RETURN, NOTE

DELETION, REPRODUCTION,

TRAVEL, COPY

EXECUTION

CANCELLATION

ORDER, (HELP)

TAB

All other

ches There are touch of

dants (O -

  a translation table which converts the codes of

  raw material in corresponding category numbers

  ) This table is in sector zero of 25354 t O file "TABLCAT-VOC" Sector one of this file contains the standard pre-WISCII translation table for pressing

  the keys It is important to note that the category table

  type is independent of keyboard positioning For example, the CANCEL and POSITIONING keys, CANCEL-

  TION, are both in the "CANCELLATION" category (13).

  This does not, however, affect the upper and lower case text characters, since they are all in

the category "; text" (7).

  The editor is divided into different operating states.

  The keys can have different meanings depending on the value of the current state, which means that a procedure table is defined for each state These procedure tables are called state tables The state tables are defined in the module

state tables.

  The voice message editor status tables

  contain indexes which designate headings of a large

  table of procedures You can find this table in the

  module of program tables, containing 36 headings.

  At the initial passage to the editor, the main state is the current operating state. When new operating states are put into operation, the old

  states, accompanied by an index of the guide-operator message

  rant on the screen, are entered in a stack of states It is assumed that the user presses the DELETE key, while he is in the main state The main state is entered in the stack of states and the definition state of

  segment now becomes current state The message guide-

  operator "What should be deleted?" appears on the screen.

  We will now assume that the user presses

  on the PASS TO NEXT PAGE key The status of

  segment is introduced into the stack, and the message

  operator guide is also introduced into the status stack.

  The new state is the passing state The message guide-opera-

  "O must we go" appears on the screen The user enters a number and presses EXECUTION This executes a

  procedure for switching to the number entered.

  At this point, the segment definition state and the operator guide message are extracted from the stack.

  "What should be deleted?" is presented again on the screen.

  The user hits EXECUTION, and a procedure is called to delete the concerned part of the voice message file. The main state is then extracted from the stack, and we

  is returned to the original operating state.

  In addition to the status tables themselves, the

  state tables also contains procedures for handling

  Read the status stack These procedures are as follows

  INIT $ STATE Initializes the state stack.

  NEW $ STATE Introduces the old state in the stack, takes

  the state specified as the current state.

  EXTR $ STATE: Extract a state from the stack, take this state as

current state.

  The status table module also contains a program which, from a category number, returns the address of the

  procedure which corresponds to this category for the state

  rant ADR $ PROGRAM Given a category, this procedure searches the current state table for the address of the procedure that corresponds to

* this category.

  The decision to call a particular procedure can thus be summarized as follows

  1) Interruption of pressing a key.

  2) GEST $ TOUCHE saves the registers, places the hardware key code in the TOUCHEBRUTE variable, calls

$ KEY OBTAINED.

  3) TOUCHEBRUTE performs the following actions a) if an unrecoverable error has occurred,

exit passage.

  b) if you hit $ PAGE POSITIONING, complete

ment of a dump.

  c) if the previous key has not yet been processed

  ted, the current key is ignored.

  d) introduction of the address of the PROCESSING $ BUTTON procedure as well as the parameter

  TOUCHRBRUTE in the program queue.

  4) PROCESSING $ TOUCHE is extracted from the pro-

  grams and executed, which accomplishes the following actions:

your:

  a) translation of pressing a key, using

sant the translation table.

  b) determination of the category number for this

  key, using the category table.

  c) if the most significant bit of the category number is zero, support for pressing a key.

  d) deletion of any error message.

  e) stop audio functions, except for

  RETURN and from the "reproduction / stop" category.

  f) call of ADR $ PROGRAM, and transmission of this procedure to the category, to obtain the address

  of the procedure we have to go through.

  g) entry into the program queue

  me from this procedure address and support

on a translated key.

  ) The appropriate program as well as the pressing of a translated key are extracted from the program queue

and executed.

  Additional procedures can be divided

  roughly in two parts There is for each struc-

  data level of the low-level modules that perform operations on this structure The file index (audio index, mark table, note table), audio functions

  and the screen are representative examples of modu-

the low level ones.

  The second part consists of the programs

  These procedures are called characteristically.

  by the key press processing system (their addresses are in the program table), and they can call the lower level programs themselves - which do most of the work.

  can therefore consider them as an interface between pro-

  management procedures for pressing the keys and procedures

  low level res doing the heavy lifting.

  The user interface module (V: voice rrr plm.

  ve user) contains high level procedures for 1.3 -

  audio, section marking and renumbering functions

ge: STOP REPRO:

INSERTION $ BRAND:

RENUMEROTAGE:

REN $ EXECUTION:

  RE NOR CANCELLATION: Procedure called each time a key is pressed in the category

  "reproduction / stop" If the function-

  audio is stopped at the relevant time, this procedure brings the cursor to the beginning of the next audio sector and starts playback If the audio function is on

  reproduction or recording courses

lying, it stops it.

  Procedure called when a key is pressed in the "mark" category If you have entered a paragraph mark, it indicates its exact position on the screen and calls the appropriate window module program to switch to this program If you have press the NOTE key, the procedure determines whether the cursor

  is on a note at the time.

  If not, it creates one. In either case, we move on to

text mode.

  Procedure called when

  a key in the category "renumbered-

  ge "l The editor is placed in the state of

  renumbering and the message guide-opera-

  "Renumbering of brands?" is present. Procedure called when you press EXECUTION when you are in the renumbering state It calls a mark table procedure to renumber the marks, to renew the information presented on the screen, and to extract

from the stack the previous state.

  Procedure called when CANCEL is pressed while in the renumbering state It is extracted from the stack

the previous state.

  The rear spacing module implements the

  backspace function Pressing the

  back space moves the cursor back five

  seconds and causes reproduction for five seconds.

  If you press N times, the cursor moves back by N * 5 seconds and playback takes place for the same time. During playback, pressing any key other than the backspace key stops playback, which

  completely cancels the backspace function.

  When the backspace key is pressed, 350 milliseconds elapse before playback begins. This is to allow the user time to press the d key repeatedly. rear spacing before reproduction begins The rear spacing module uses three variables to perform these functions: mode: time $ ea: count $ reprod $ ea: State "1" if spacing is performed

  rear, "O" otherwise.

  The time corresponding to the cursor when the user pressed for the first time

  times on REAR SPACE The reproduc-

  tion will take place up to this position ,.

  but not beyond, regardless of the number

of pressing this key.

  Counter decremented by the RYTHIEUR $ DIX $ MS procedure Used to count the

350 ms waiting time. The backspace function provides the procedures

hard-following: EA:

COUNTER $ WAITING $ EA:

CONTR $ KEY $ EA:

  Procedure called when the backspace key is pressed If

  support is the first, she positions-

  does $ ea mode at "1" and recalls time $ ea.

  It initializes time $ waiting Sea to

350 ms.

  Procedure called every 10 ms by RYTHMEUR $ TEN $ MS This procedure decrements time $ wait $ ea, and when 350 ms have elapsed, it enters the queue of

  programs a procedure that causes

  ra reproduction from the current cursor position until time $ ea. Procedure called by TREATMENT $ TOUCIIE This procedure cancels the backspace mode if a key other than the

backspace key.

  The cursor module has all the high-level cursor functions. Again, these procedures are simply interfaces between key and screen processing programs that actually move the cursor over the screen.

PROG $ CURSOR:

PROG $ PASSAGE:

PASSAGE $ EXIT:

$ CURSOR PASSAGE:

PASSAGE $ ACCEPT $ NUMBER:

EXECUTION PASSAGES:

  Procedure called in most of the states when a key is pressed in the "cursor" category It simply calls one of four screen programs, depending on the cursor key on which

supported.

  Procedure called when pressed

  on the SKIP TO NEXT PAGE key

  VANTE It introduces the old state into the stack and ensures that

  the current state is the "Passage" state.

  It presents the message guide-opera-

  "Where should we go?" and move the cursor to the next position

  immediately the message guide-opera-

Note that at the time of the

  translation of the message file, this-

message operator guide must be

framed on the right.

  This procedure is called when you press CANCEL while in the "Passage" state. It brings the cursor back to the audio / mark part of the screen and extracts the previous state from the stack. Procedure called when one of the cursor keys is pressed while in the "Passage" state It calls one of four screen programs depending on the cursor key which has been pressed It then calls PASSAGES $ ORTIE for

return to the previous state.

  Procedure called when a key is pressed in the "number" category,

  while in the "Passage" state.

  This procedure displays the number on

  the screen just after the prompt

  operator and updates the cursor position. Procedure called when you press EXECUTION while in the "Passage" state If a number is present on the screen, it is converted from ASCII to binary and a screen program is called to position the cursor

  below the appropriate mark.

Then she calls PASSAGE $ EXIT

  to return to the previous state.

  The text entry module contains programs for entering text notes, in the text mode. The following variables are used: buffer $ text (60) indext cursort index $ note

first-

  stamp to keep the note of

text during its introduction.

  current position (0-59) in the tam-

text pon.

  current cursor position on the screen. index designating the note table of the

  text note we are working on

at the time considered.

  indicator, in state 111 "if the note which has been entered has just been created In this case, pressing CANCEL will delete this note If it is an old note which is

  modified, press CANCEL

  will simply blur the note to its original form. The following programs are provided:

TEXT $ FIXED $ FIRST:

ENTER $ MODE $ TEXT:

CANCEL $ TEXT:

EXECUT $ TEXT:

  Procedure called by INSERT $ BRAND

  to indicate to the following input module

before this note just comes

to be introduced.

  Procedure called by INSERT $ BRAND

  when the NOTE button is pressed.

  Introduces the old state in the stack and fixes the new state "Text" Brings up the operator guide message "Enter the text" Takes the note in the note table and places it in

the text buffer.

  Procedure called when you press CANCEL while in the state

  "Text" If you have introduced a new-

note, this note is deleted.

  Otherwise, the text buffer is rejected and the screen shows

  new the old note unchanged.

Restores the previous state.

  Procedure called when you press EXECUTION while in the "Text" state Replaces the old note with

  the content of the text buffer Reta-

blit the previous state -

CURSOR $ TEXT:

$ BACK SPACE $ TEXT

ENTRY $ TEXT:

TEXT:

  Procedure called when a cursor key is pressed while in the "Text" state Moves the

  cursor forward or backward Affi-

  che an error message if you press

on Curseur Nord or Curseur Sud.

  Procedure called when the backspace key is pressed

  while in the "Text" state.

  Move the cursor back one position, then delete the character under which

is the cursor.

  Procedure called when a key is pressed in the categories "text", "number" or "reproduction / stop" Enter the character in the text buffer and on the screen and

advances the cursor one position

  tion. Procedure called when a text key is pressed while in the "Main" state If the

  cursor is on a note, this process

  hard switch to text mode and enter the information of the key pressed in the text buffer and on the screen If the cursor is not on a note, this procedure presents

the message "Move the cursor".

  The editing module interfaces the key processing system with the low-level, screen-related file index programs that accomplish

  actually the manipulations on the file.

  The editing module keeps track of the parts of the file that are edited We use a point structure to designate positions in the file This structure is of the form point structure (address time, multiplet index) in which: "time" is the elapsed time in the file, and "index" is the mark index of the current mark, or of the following mark in the file if there is no mark at this position. The following point structures are used to keep track of positions during editing: pointstart the beginning of a segment to delete / move / copy pointend the end of a segment to delete / move / copy point is the destination point for a move / copy operation. To delete part of the file, the segment between start point and end point (included) is removed from the file. To move or copy part of the file, move or copy the segment between start point and

  pointfin (included), to bring it to pointdest.

  When inserting into the file, pointdest becomes the insertion point The current end of the file is startpoint, and recording begins at

end of file.

  When the user presses STOP the program performs a movement, as described above, by moving towards pointdest the segment which is delimited by (startpoint, endpoint). To replace a segment of the file, three additional point structures are used:

  rpointstart contains the start of the segment to be deleted.

  rpointfin contains the end of the segment to be deleted.

  rpointdest contains the start of the segment to be inserted.

  The replacement procedure takes place from the

  as follows: we initially define the segment to be

  place between pointstart and pointfin Once the segment is defined, we copy pointstart to rpointdest, and pointpoint

  to rpointfin, and we set rpointstart at the end of the file.

  We then carry out the classic insertion procedure, saving at the end of the file As with insertion, when we press STOP, the new material, i.e. the segment (start point, end point) is moved to the point

  insertion point, i.e. pointdest, to complete the insertion.

  During the replacement, the user can insert, repro-

  mute, press the cursor keys, and - enter paragraph marks and text notes All

  insertions are accomplished in the normal manner, using

  sant pointdébut, pointfin and pointdest Of course, all insertions are limited to the region beyond rpointdébut.

  If the user presses CANCEL, replace it

  ment is canceled by replacing the final point in the voice file at rpointstart, which restores the file to its original form.

  If the user presses EXECUTION, replace it

  ment is executed by first deleting the segment (rpointdest, rpointfin), then assigning rpointdest to

  pointdest, and the end of the file at pointfin, then in effect

  killing insertion using normal displacement of the

  segment (startpoint, endpoint) to pointdest.

  The audio function module contains programs

  my intended for reproduction and recording in voice files It uses an associated module, the input / output module, which contains data structures and procedures for handling buffers and requests for

  queue directed to the master module.

  During playback or recording, audio data must be recorded in a buffer so that

  that reproduction or recording is not interrupted

  pu by waiting for the completion of a write operation

  or reading in a buffer Workstation software

  vail audio is designed to use at least two

  pons, but we can use more if space allows

  At present, the audio workstation uses

6 audio buffers.

  The voice message editor uses buffers that are from 1 to 16 sectors in length These buffers are aligned per page in memory Each buffer corresponds to an audio block in the voice file The input / output module Contains structures, called structures that manage the audio buffers The input / output module contains

  a queue of input / output requests, which we use

  se to organize the blocks (RCB) in queue The 10 ms timer controls this queue every ms If it contains anything, the rhythmer's procedure extracts the request from the queue itself

and presents it to the master module.

  The input / output request queue used

  reads the following data structures: queue an address table, constituting the

  input / output request queue

tie.

  high index from the top of the queue.

  bottom index of the bottom of the queue.

  counts number of items in the queue.

  The following programs handle the queue.

  you. INTRO From ES Introduces the address of an RCB into the queue for input / output requests. EXTRACT $ AND $ SEND: If there is something in the queue and if the SCA is free, this procedure extracts the RCB address from the queue and places it in the SCA This procedure is called as soon as we put something in the queue (she tries to check it out immediately) - She is also

  called every 10 ms by the process

lasts RHYTHMUR $ TEN $ MS.

  Because the voice message editor only inserts recorded data, it does not perform overlay recording, and recording always begins at the end of the file. The inserted data is saved when

  end of the file and are then moved to the point of in-

setting.

  To record, the following operations are carried out:

touts

  1) start with the sixth information structure.

  a) fill in the first buffer address b) fill in the buffer size c) if you save in the last block of the file,

set the stop indicator.

  2) indicate to the equipment the address of the first stamp.

  3) tell the equipment to start recording.

  4) complete the following procedure: a) indicate to the equipment the size of the buffer in which

  it records at the time considered.

  b) enter a write request in the queue

  for the previous buffer, if it is not the first-

mier tampon.

  c) if the stop indicator is set for this tam-

pon, stop.

  d) determine whether a previous write request for

* this buffer has been satisfied; if not, stop

  ter the audio function until the request has

been satisfied.

e) fill in the RCB for this buffer.

  f) increment the -variables so as to be ready to

process the next buffer.

  Once the hardware has completed recording to the first buffer, a block account interruption

  is produced (channel CTC 0) When this occurs, the procedure

  duration INTERRUPT $ AUDIO is called This procedure performs a

  check to see if the mode of reproduction or recording

  very much in operation, and it calls a procedure

  of interruption of reproduction or recording The opera-

  tion 4), above, is the procedure for interrupting recording

  REGRUPT $ RECORDING While the recording is

  sticks, it is called each time a tampon is finished.

  Reproduction is similar to recording.

  We carry out some initialization operations, then we indicate to the material to start reproduction We immediately call the program INTERRUPT $ REPROD During the reproduction of each buffer, we call again INTERRUPT $ REPROD to prepare the next buffer for reproduction, and for queue a request for

  read another buffer from disk.

  During recording, the sample rate-

  swimming is always set to the value corresponding to the

  ral CAD $ SAMPLE, which defines the sampling rate.

  However, during playback, the sampling rate can be changed Every 10 ms, the RYTHMOR $ TEN $ MS procedure calls the FIXATION $ CADENCE procedure This procedure calls a program intended to convert the current setting of the speed control to give the appropriate sampling rate The value of this rate

  sampling is then transmitted to the equipment.

  The voice message editor screen is divided into

  two sections: the status section and the audio / market section

  ques The status section is formed by the first two lines and by the last line of the screen. We use this area to display operator-operator messages, the time corresponding to the cursor, the length, etc. We use the audio / marks, which is formed by lines 3 to 21, to present the -content of the voice file, i.e.

  audio blocks, text notes and paragraph marks

  phe. The display module controls the status section of the screen. This module also contains all the

  GROUPEMENU procedures It contains procedures for initi

  tialize GROUPEMENU, and to view the corresponding time

  with the cursor, the audio mode, a memory aid; the mode of

  telephone, title, operator guide messages, length

gueur and error messages.

  The window module contains the programs for

  born when viewing and updating the game

  audio / screen brands This module is assisted by the modu-

  the following: conversion (V: voice rrr plm ve conversion) Time position structure conversion programs (V: voice rrr plm ve time)

time conversion

  line position (V: voice rrr plm ve line) Implementation of the line structure

  region (V: voice rrr plm ve region) Information searcher

  scrolling editing dex (V: voice rrr plm ve scrolling) Low level window operations $ The voice file includes a header, a table of

  marks, a table of notes, a plan of establishment of the sec-

  blocks and a block layout plan The following modules

  vants contain a program allowing access to the voice file: file index (V: voice rrr plm ve index file) Realization of editindex mark note voicegrm extension the index of fiecrih (V: voice rrr plm ve editindex) Editing operations on the file index (V: voice rrr plm ve mark) Realization of

the table of mar-

  ques (V: voice rrr plm ve note) Realization of

the table of

  notes (V: voice rrr plm ve voicegrm) Programs of

creation, ini-

- * tialization and

of implementation

  vre of the voice file (V: voice rrr plm ve extension) Extension and truncation programs of the voice file

  irrecup unrecoverable error, pro-

  ABEND management gram The Error Module contains procedures for ABEND programs, concerning unrecoverable and recoverable errors An indicator, INDVIDAGE, set in the group, is used to determine whether or not an error leads to a dump If INDVIDAGE is OF Fh, emptying is validated If its

  value is 0, dumps -are disabled.

  The procedures provided are as follows:

  ERROR $ RECUP: Emptying if the indicator is set, display

INFORM $ ERROR:

ERROR $ IRRECUP

  error error VE: XXX, designating by XXX a transmitted error number These numbers

  are defined in (V: voice rrr lit.

  ve.ERR) There is also display of a data section with 16 bytes (so

  characteristic of an RCB) in the event of transmission

a parameter.

  Display of an error message not VE, return to the calling application, after

  press any key

  non VE error errors are simply normal errors such as "Move cursor" which are displayed at the bottom of the screen These error messages

  are defined in (V: vo rrr lit ve.

HAPPY).

  Same as ERROR $ RECUP, except for

  fact that this error is not recoverable.

  The editor returns to the calling program when the user presses any key. The message editor recovery system

  voice allows recovery from power cuts

  workstation or accidental initial program load (IPL) operations during the process

  The voice message editor uses some

  our current data structures, and three modules

  hold these structures and programs to manipulate them.

  The program queue uses the procedures

following hard:

  INIT $ QUEUE This procedure defines a queue.

  The user specifies the address of the queue,

  the size of the queue, the size of each element

  lie in the queue and a pointer to a

  structure which contains all the characteristics

  important ticks in the queue This structure identifies the queue It must be passed in the form of a parameter to the introduction and extraction programs - which

are described below.

  INTR $ QUEUE: This procedure introduces an element into a

specified queue.

  EXTR $ QUEUE This procedure extracts an element from the head

of a specified queue.

  The stack module (V: voice rrr plm ve stack) is an embodiment of a stack with input and extraction programs The stack of status table modules uses procedures coming from the stack module to perform the battery

  Unlike the queue module, the programs

  grams of the stack module can only work on a single stack, defined in the module as follows

  stack (12) space (in bytes) reserved for the stack.

pp stack pointer.

Two programs manipulate the stack

  INTRODUCTION Introduces an element into the stack.

  EXTRACTION: Extracts an element from the stack.

  The bit table module (V: voice rrr plm ve bit) can position, restore and test bits in a user-specified bit table The table cannot

  not contain more than 256 bytes The mark table used

  read a bit table to determine the number of the mark, next paragraph to create The file index editing module uses a bit table to order all the free blocks in the index, so that the extensions The file table module contains the following procedures:

  POS $ BIT Sets a bit in a bit table.

  REST $ BIT Restores a bit in a bit table.

  TEST $ BIT Tests a bit to see if it is set or restored. All PLM input and output instructions

  for the voice message editor are contained in the modu-

  the audio equipment control (V: voice rrr plm ve comau-

  dio) This module contains small procedures which make

  interface function between the equipment and the main part

  blade of the voice message editor PLM code.

  The fixed interrupt mode module (V: voice rrr z 80 ve modintfixed) contains two procedures,

  one to place the workstation in inter-

  ruption 2 and the other to bring it back to interrupt mode 0 The PLM, INIT $ POSTWORK and

  RESTAUR $ POSTETRAVAIL, which can be found in the module

  request audio material, call the two programs located in the fixed interrupt mode module The very first bytes of this module contain the interrupt vector tables for the CTC and the PIO (input / output management program for the These tables must reside in memory within limits corresponding to a factor of eight, so care must be taken to verify that

  this is done in the link table.

  It goes without saying that many modifications can be made to the device described and shown,

  without departing from the scope of the invention.

Claims (6)

CLAIMS -   1 information processing system, characterized in that it comprises: continuous signal acquisition means (54) intended to acquire a continuously varying electrical signal which corresponds to a voice message; digitizing means (14) for digitizing the continuously variable electrical signal to produce discrete voice data corresponding to the audible quality of the voice message; discrete data acquisition means (18) for acquiring discrete data corresponding to alphanumeric characters; means of acquisition   discrete signals (20) for acquiring discrete signals   crets comprising editing and ordering orders; a   memory (22) for storing data in form   discreet me; display means (31) for creating a display; and a processor (26); ways   continuous signal acquisition (54), the digitization means   tion (14), the discrete data acquisition means (18), the discrete signal acquisition means (20), the   memory (22), the display means (31) and the process   sor (26) being functionally interconnected by connectors   control conductors and data transfer channels (58, 60, 62); and in that an operating program for the processor is stored in the memory (22) so that the processor (26) controls the operation of the system for: recording the discrete voice data in the memory (22), simultaneously with the 'acquisition of the voice message; save the character data in the memory (22) simultaneously with the character input, establish in the memory (22) a sequence record indicating an order   unified voice message and character data, visualized   ser a sequence of voice symbol marks and marks   of characters, each symbol mark representing an incr   predetermined wording of the voice message acquired and each character mark corresponding to one of the characters entered, the   sequence displayed corresponding to the sequence in the   record, and revise the sequence record, subject to input edit commands, to take account of edit changes in the order of voice and character data. 2 System according to claim 1, characterized in that the operating program is designed so that the processor (26) also controls the operation of the system for: reacting to predetermined discrete signals acquired simultaneously with the acquisition of the voice message, to indicate in the sequence recording the instant at which each of the predetermined discrete signals was acquired, and present in the display a distinctive indication of the instant of acquisition of each signal acquired simultaneously,   compared to other elements of voice data.   3 System according to any one of the claims.   1 and 2, characterized in that the function program-   This arrangement is designed so that the processor (26) also controls the operation of the system for: establishing in the memory (22) a pointer defining a pointer position in the data sequence, presenting on the display a visible mark corresponding to the pointer position, and move the pointer position which is defined in the sequence, and move this position correspondingly in the visualization, depending on signals input acquired.   4 System according to any one of the claims   1 to 3, characterized in that the function program-   It is designed so that the processor (26) further controls the operation of the system to produce, under the dependence of acquired input signals, a continuously varying audio signal which corresponds to the discrete voice data recorded in the memory (22 ), the production of this signal starting at a point in the data sequence 25354 O   which corresponds to the defined pointer position, as defined at the time in question, and following   the order which is defined at the time considered in the recording very sequence.   5 System according to any one of the claims   1 to 4, characterized in that the operating program   operation is designed so that the processor (26) controls   furthermore the functioning of the system to advance   the pointer in the voice message data, correspondingly   dance with the progress of the audio signal production.   6 System according to any one of the claims.   1 to 5, characterized in that it comprises a circuit (28) intended to detect an audio signal acquisition activity and to interrupt the recording of the data of   voice message in memory (22) in the absence of activity.